A hallmark of a well-functioning insurance market is protection against the risk of catastrophic financial losses. Historically, protection from large losses resulting from mental health/substance abuse (MH/SA) service use has been unavailable in most private health insurance plans because of special benefit limits on MH/SA services. Parity policies aim to protect individuals with severe mental illnesses against the catastrophic costs of seeking treatment, and, in doing so, increase efficiency and fairness in the insurance market. In 2001, all plans participating in the Federal Employees Health Benefits (FEHB) Program were required to offer coverage for MH/SA services on a par with general medical services. The FEHB parity directive represents the most comprehensive parity policy enacted to date. Previous studies of parity have found that, on average, implementation of MH/SA parity did not result in either the large increases in MH/SA spending predicted by opponents or increased access to services anticipated by mental health advocates. Parity has resulted in relatively small reductions in average out-of-pocket spending for service users. It is possible that the relatively small changes in average out-of-pocket spending resulting from parity adoption were driven by the subset of individuals with the highest costs;however, this issue has not been examined. The impact of parity on the sickest individuals and those carrying the greatest financial burden of illness constitutes the central question left unanswered by the existing literature on parity. The specific aims of this application are to: 1) Compare the effects of comprehensive parity on total and out-of-pocket MH/SA spending across the distribution of expenditures in order to determine whether parity differentially increased financial protection for individuals with high MH/SA expenditures;2) Examine the effect of comprehensive parity on the composition and intensity of the treatment they receive;and 3) Assess the effect of parity on the likelihood of being rehospitalized for an MH/SA condition and the likelihood that patients diagnosed with bipolar disorder or major depression receive appropriate ambulatory treatment. We will use a quasi-experimental design to compare FEHB plan enrollees with a matched comparison group of privately-insured enrollees not subject to the FEHB parity directive to account for secular trends in MH/SA use and spending. We will attempt to disentangle the effects of parity and managed care carve-outs by comparing plans that used carve-outs with those that did not. We will estimate a combination of two-part random-effects models for longitudinal data, latent class random-effects models, logistic regression models, Poisson regression models, and hazard models to address these aims.